Google Research describes AI agents for wet-lab discovery

Google Research said it is bringing AI agents into wet labs, with a system meant to connect high-level agent reasoning to laboratory robotics and the physical limits of a bench setting. The project description says the work is aimed at closed-loop molecular design cycles and protein binder discovery for cancer targets.

The company said one of the pieces is VERID, short for VERified Infrastructure for Discovery Laboratories, which it described as a semantic digital twin of a laboratory workcell. According to the description, the framework tracks robotic arm geometry, well volumes, reagent identities and pipette tip inventory.

Google also said the effort is meant to support safer and more reliable autonomous discovery, including checks on whether agents are optimizing for real target binding, verification of multi-step wet-lab protocols and monitoring of reasoning and execution traces. The company described the work as building blocks for human-directed autonomous discovery in the life sciences.

Sources

• Grounding AI agents in biological reality: Building blocks for autonomous wet labs